Skip to main content

About this Show

Book TV After Words

Sasha Issenberg Education. (2012) 'The Victory Lab The Secret Science of Winning Campaigns.'

NETWORK

DURATION
01:00:00

RATING

SCANNED IN
San Francisco, CA, USA

SOURCE
Comcast Cable

TUNER
Channel 91 (627 MHz)

VIDEO CODEC
mpeg2video

AUDIO CODEC
ac3

PIXEL WIDTH
704

PIXEL HEIGHT
480

TOPIC FREQUENCY

Obama 16, Us 6, Sasha 2, Chicago 2, North Carolina 2, Offline 2, Michigan 2, Augusta 1, Boston 1, Academia 1, Philadelphia 1, Academe 1, Propublica 1, Rick Perry 1, Gingrich 1, Perry 1, Bush 1, Emily 1, Keeshan 1, Wright 1,
Borrow a DVD
of this show
  CSPAN    Book TV After Words    Sasha Issenberg  Education.  (2012) 'The  
   Victory Lab The Secret Science of Winning Campaigns.'  

    October 7, 2012
    11:00 - 12:00pm EDT  

11:00am
i'm not sure during the census in 2010 from wiki and the population of almost 550 people, which doesn't sound like a lot, but was about 8% to 9%. was the first time against augusta route. with the downtown area that is revitalizing and growing businesses. ..
11:01am
>> host: sasha, this is just such a timely and provocative book, as we know where weeks away from this election. i just wanted to know, how did you come to want to write this book? >> guest: i cover campaigns begin in philadelphia, and so i always pay more attention to sort of passive technique just because i'm from a big city, so much attention was being paid to the vote counting and precinct targeting. so i talk to more people, and i was always shocked as a think anybody who's spent a lot of time run campaigns is that most of the people i talk you could explain to me why they did anything that they would do. like how do you know that, how do you do that? at some point they did because that some sort of rule that was really based on any research.
11:02am
and so i sort of when run campaigns to some degree with skepticism, the practices that were taking place and the way they were spending time, and as big as i learned about people, starting in academe who are doing the field experiments, randomized control trials, within being adopted by people in the political world, and fund more about all the innovations of data, targeting based on, basically revolutionize campaigns in the last decade. this was a major generational shift in that in addition to all the new forms of research changing, the way campaigns operate issue that this kind of cultural tension between a lot of the old practices and the sort of new empirical movement. i cover the 2000 campaign -- 2008 campaign for "the boston globe" and writing mostly about the message stuff going back and
11:03am
forth. and after 2000 i started to burn down in this burgeoning world of people doing empirical research, and was so convinced that that's where the untold story of a lot of what's really going on in elections these days is taking place. >> host: interesting. what you want the reader of this book to most understand? when they read this book, what's that nugget you want them to come away with? >> guest: i want them to understand that what we see in tv and in the newspaper every day, the content of the campaign is very little of everything that is going on. you know, that campaign are far more than once said in their ads, some other ads, with the candidate goes, and it's running mate or spouse goes. and i think, and what the spokesperson says on tv, i think too much of a campaign coverage basically assumes that this is
11:04am
sort of the entirety of the campaigns activity. most campaigns and especially to a presidential level, most of the people are saying that campaign is doing other things. that's surfacing that we see, a candidate giving a speech, buying ads something to go on tv, in many ways hasn't changed in decades. stuff that a try to write about in this book that's taking place with people don't go on tv, almost entirely unknown, the outside world has changed dramatically is changing, didn't come every year, sometimes every month. campaigns are getting a lot smarter about what they do, why they do, how they do it. and i think that in general campaign coverage is sort of struggling to keep up explaining to the voters, the viewers what they are doing and why and start thinking the sum of undergirding all that. >> host: i'm just thinking it's called the victory lap,
11:05am
sort of you have his picture in your mind that it is this lap. can you give us some perspective in these campaigns now how many people are working on this stuff? if you can tell us what is a typical campaign, how many people are usually working on this campaign and how many people would be working on this sort of data, analysis and all things you're talking about. >> guest: on some level on smart campaigns and data analysis, pure system is sort of informing everything to do. there's more and more stuff on the campaign to contest that either through rigorous experimental methods or through some sort of come even if it's not randomized, a lot of these extremes are. there some sort of discipline testing. so smart campaign, and the obama campaign is symptomatic of this, basically thinking everything the campaign does is form by david. and you get down to the state
11:06am
level campaigns and they definitely are having people who are voter file managers are basically dealing with data or talking people are just dealing with david. you get up to presidential campaign and one of the things the romney campaign had to do this summer after they won the nomination was what they called build a data science team. increasingly that sort of function is becoming a core function of the campaign. used to be to the extension of data it was left for fund-raising or you could buy vendors our consultants. and now, you know, people will have call them different things but there's basically the core function of a modern campaign to people, especially on the photo site, just crushing and processing data. >> if any of us were to go to the romney campaign or the obama campaign and where to look around the headquarters, how many people, is there a lot of
11:07am
young staff? what does it look like? >> guest: chicago, dozens of people depending i had how you define it, analytics him and then in every state they are hiring for jobs that are dated jobs, voter file managers, targeting directors, that's, you, the obama campaign will have thousands around the country and i guess hundreds of them are directly interacting with the data everyday. >> host: do you think one of the parties, republicans or the democrats, is more adept at using this technique? or are they all sort of at the same level? >> guest: i think at the moment democrats are ahead. there's some leapfrogging cycle to cycle with different parties. and one of the things that becomes clear as i looked back over where innovation takes place in politics, i think it's one of the stories, these campaigns are really -- utility company, it lasts six months or
11:08am
18 months. you don't really always know what your budget will be in defense but you're concern about market share on monday. nobody really has an incentive to invest in any research which will yield a day after election. typically the person been chairman of the board of the campaign of the candidate, there's two types of them i think of it. someone has done it before and thinks they know how to do, maybe has been doing over and over again and just perfectly happy with keeping things the way the worker or just someone who's never done before and this was somebody who would want to be a senator or a governor or sheriff doesn't how to run a political campaign. the more technically advanced campaigns are getting, the less candidates should be expected to sort of understand the mechanics of these things. and so, so campaigns usually are not places that are built to innovate and are not built to learn from their failures. so i look back over a special
11:09am
presidential campaigns using our reelection campaigns carve certain institution to they exist more or less for four years in some form or another. they been able to plant, budget over four years, stepping over four years. they often don't have primaries that distract them in the short term. and for that reason the bush '04 campaign was revolutionary and using some microtargeting tactics of plotting them, and the obama campaign 2012 is similarly a place where sort of exceptional innovation and other techniques is taking place. and part of that is that you can set out a research agenda. and in chicago, postmortem committee that were set up in november 2008 for people doing data targeting, put together reports almost four years ago now, been certified to what they did right, what they did wrong
11:10am
and fix the future. it's really unusual. or even parties where there's not confusion, usually that much continuity. in this case will have come it will be a lot more than a million dollar come by to answer what can we take on. so that culture, that institution building that's taken place around the obama campaign i think has helped give the democrats a major event. in the '08 campaign as i write about in the book, there is an exceptional experience there, partly because of the long primary season and a huge staff that they had that somebody people on the campaign, at junior levels came into the contest was pretty sophisticated tools and analytics. a lot of those kids who were a field director using advanced voter data in 2008 went out and went to state party or labor unions, other institutions on the left, and there's a sort of pollination that is taking place
11:11am
that i think a little more active on the democratic side right now. >> host: so you would give probably obama the advantage with this sort of type of technique? >> guest: i think so. the romney campaign has proven itself to be very smart, disciplined and i think a number of their biggest successes sort how they went about winning iowa and without growing too much attention to it for a long time and how they won subsequent states, all of those pretty small footprint on the ground were evidence of their ability to outfox gingrich and other opponents. by using data. but they are scrambling to sort about anything approaching approaching the scale of the obama campaign. anything approaching the ambition of the obama campaign, the type of policy to be solved
11:12am
in six months are going to be a lot smaller than the type of problems you can solve in four years. >> host: do you think that this campaign is going to rely more on messaging his presidential campaign, or on these targeting techniques? what's the next? how much -- test message still matter or is it if you can just fine-tune it and target a better, that's going to be the difference? how do you put that altogether? >> guest: the message still matters. one of the things, you know, this whole narrative, the bulk of my book takes place over the last 10 to 15 years, has taken place in an era of real partisan polarization. one of -- 2000 that being his formative moment i think in terms of innovation in politics, for two reasons. one is it's a very close election. so campaigns that might have once said i get 1% are 2% of vote if i use this technique, after florida people said i
11:13am
could use one or 2%. so they're starting to be a real focus on techniques that can provide very small measurable boost. that's one thing that happened after 2000 or the other thing was we realize we are moving at this really deadlocked polarized era, maxed out. started writing a memo by the supreme court was still taking up the case, an and in it he lod at polling data from 1984, something like 25% of americans were party switches, mood between the two parties. and by 2000 i think it was 7%. and we are basically in the era now where a very small percentage of population is actually persuadable between each party in a general election. most voters are predictable. they may not -- there are far
11:14am
fewer voters who are actively moving between the two parties that were previously. so when you get into that environment, excellent issue for campaigns to visualize where they can get benefits by focusing on turnout or registration of their supporters as opposed to merely trying to persuade them. now, this you i don't know if it's 8% are not present that are persuadable. the campaigns will focus on them. but we have a far better science now in understanding what motivates people to vote. and a lot of it -- the science of mobilization turnout has gotten much better. the science of persuasion, it's still pretty vague. and so i do think that there's been a sort of reinvesting in a lot of mobilization techniques
11:15am
in part because of what we've learned in the last decade. sunday of these two separate things. you certainly know what you give to somebody what you can do to increase their likelihood of voting. and now have better targeting techniques through data to forget who you talk to and about what. and so i don't think of it as necessarily message or targeting, but good campaigns do targeting and analysis on the front and that allows them to understand in a far more precise clean way for their turnout targets but they don't need to talk to until it's time to push them to vote and who the persuasion targets are. then if you're narrowing your universities are trying to persuade you to make your messages a lot sharper. you can sort of focus your reached, research and focus groups and then pulling, and an experimental testing to get more closely to the question of what it's either 7%, talking to 7% who are really persuadable under
11:16am
not trying to come up with sort of messages that are speaking to a far broader electorate and trying to persuade them all at once. >> host: i know in the book you mentioned different techniques that are used to actually get people to vote. i know this country is such a low-level many times it sort of participation but can you explain some of these techniques that are used to get people voting? >> guest: the best ever demonstrated tool for getting people to turn out and vote was measured in its experiment in michigan in 2006 before the republican primary. and this consultant vendor named mark had sort of the practice of threatening to expose people if they didn't vote. and he organized a experiment with these two political scientists at yale who have been pioneers in using basically the same message that people use for drug trials to campaign techniques. randomized control trials. so instead of distributing a
11:17am
pharmaceutical company could send different types of mail or phone calls in and go after those who voted and who did and see what has affected for the first time to able then to really disentangle cause and effect in particular political tools. >> and in michigan they found that remarkably effective with sending post a copy together before the election that said dear nancy, as you may know your history as a voter is a publicly available docket on file. as you may know there's an election coming up in a few weeks. after the election we will send you and your neighbors a copy of your vote history. and it had a line with, the recipients vote history like the last six elections. didn't have a bunch of people
11:18am
with addresses on the same block and said we'll senate updated set of these. he got death threats when these went out. tons of other sort of schooling for people who felt it was bowling. but it turned out to be remarkably effective, what psychologists call social pressure. the expectation that people would have, friends and neighbors now not have a vote on which candidate for for but whether they actually did go out and participate. and so this paper was published and it was a revolutionary thing because it basically was like 10 times more, sometimes were affected by anything they had have been used before to mobilize people to vote. no campaigns, parties, institutions wanted to put their name on anything like this because it really did seem like some form of blackmail and it was a real fear of backlash. but over the next four or five
11:19am
years a lot of institutions, many of them on the left, academics working with unions, the dnc, candidates, tried to experiment with how to soften the language. and so now with sort of widespread practice of sending a letter out that they called thank you treatment, and it says something like dear nancy, thank you, publicly federal records show utah in the last election i want to thank you for being a voter. there's another election, up and i'm hoping to ask for a second thank you again for participating. people have done these over and over again. there's nowhere near as effective as a sort of threat to expose people. the behavior psychologist who work in this space gets at the same mechanism, which is that i think that voters have become so conditioned to the idea of a secret ballot, many of them don't know that their vote history is publicly available to every campaign knows how often you vote in an election.
11:20am
and don't realize that they are sort of, that they're visible in a sense. and that they can be monitored. and so this idea of hinting at people that there's some form of surveillance i can take place and the people will judge you, you know, it's -- people want to be seen as voters of this state are going to get them to be as voters, these have shown time and again they would. >> host: i guess is anybody use that in the primaries? i know that's the place where people are not showing up to vote. is that a technique? >> guest: certainly around the country people have. i don't think in this book and primary i've heard anybody using that, but it's become so widespread in certain areas that people are not afraid it sort of lost it, its potency, that has no element of surprise or voters when they get a piece of mail that kind of nudges them that way.
11:21am
once voters have sold of a culture the idea they're being surveilled, then there's only so much you can get in conditioning their behavior the next time. there's a sort of constant scribbled upon the things that's taken place in the spirit i read about in the book is behavioral psychology is informed so many aspects of human life in the last generation or so, certainly economics has been transformed by behavioral insight. politics, clinical science has been late to accept them. and once local scientists sort of rediscovered the use of experiments to go out in the real world and test techniques and campaigns on voters, they sort of naturally went to basically want to be oversize textbooks and said what of the things people using to motivate individuals to recycle or to buy certain products? change their investing habits,
11:22am
and that's if we translate these things to voting for registering. so there still is i think a lot of stuff that's been demonstrated in realms other than politics of having a sort of psychological mechanism that people work in politics think can be translated into come into political behavior. and so it's almost all the motivation, not isolation. but i think there's probably going to be far more that comes out of that world that changes the language and texture a lot of the political teen occasion that we have. >> host: do you think most of this right now is more weighted towards motivating people to vote, is that what you're saying? and less really figure out the sides between the persuasion? >> guest: part of it is the way people to sign these experiment. the easiest thing to measures whether not someone voted. you go to the board of elections a month after the election, they've updated the voter file. it's yes or no, they voted. persuasion, to see if they
11:23am
change their mind, have to rely on them being honest, their self-reported other choices. so, and the other reason a lot of the people who start doing this work in academia were using basically nonprofit dollars through the institutions and couldn't do partisan work which made it very difficult to do. persuasion in the campaign, but if you're doing non-candidate specific test you can spend university dollars out of your research budget on a. there's a big body of work on the science, and a lot of it is informed by the behavioral psychology. there's far less sort of i think new science of persuasion. one thing i've written about the issues how the obama campaign is trying to adopt these as primitive methods to measure the effect of their mail and the online ads and tv ads and just as propublica, more expensive.
11:24am
but if you would tried to break out of relying solely on focus groups that try to ask voters to imagine any sort of artificial setting that would change the mind and instead introduced into information in the real world where they don't believe they're basically being examined. and see if it accidents change their mind. so that's sort of one of the next was i think of using these experimental tech makes to see what really does move voters and get out of sort of the artificial setting of the focus groups on loan. >> host: so on that note with the obama campaign since they've had four years ago time to test it and now they have four years later, is it any innovative or different techniques that are using now anymore that you could -- >> guest: these programs are really the thing. they have been used in small airways by some institutions,
11:25am
afl-cio, emily's list over the years your in certain states, but using them in a presidential campaign as he obama campaign has this year is a major advance. part of the problem with the way that people have tested messages historically is that you either have a poll that asked some of who they support, and it tells them a bunch of defamation and to ask. >> into sport and you see anybody those. before you ask directly if i told you that mitt romney hasn't paid taxes for x number of years, would it make you more or less likely to vote for him? some% of people either move when they hear that a mesh or they tell you they wouldn't. people telling you that they would move is a pretty sort of conjectural thing. i wouldn't cognitive growth anyone who told you what they would do under hypothetical senate and given some information. they may or may not already
11:26am
know. that's part of the problem with those polls is that asked people if new piece of -- you may or may not already know. and so, or in focus groups you bring somebody in an jafa doesn't people and show them and add and ask them again, did anyone change their mind. and you're pumping somebody to change their mind. and they're being forced to watch an ad that they might otherwise tonight. so using these experiments programs of the obama campaign has, introduce them into the real world. they randomly assigned mail randomly to pay part of the electorate or the assigned tv ads to the mark. been because they're pulling across those markets vacancy who moved base on which message of which type of ad army. and the because if all this data, thousands of individual data points on either -- each vote they can look at the attitudes of the people who did not.
11:27am
not just people who said they would move. and so you can start to statistical model the traits of somebody who moved when they see an ad about mitt romney's taxes. instead of just pulling people and looking at the time for people who said they would move. and so then they certified down to a cycle of a few weeks to develop tests, randomly assign and distribute these ads and to measure the movement. and so at the end of it you realize okay, the people who are moving on the tax message, let's say, or can conservative retirees, okay, now let's go send these ads to conservative retirees. and that's a radical new way of both i think transferring some of these techniques to measure persuasion, and also getting rid of all the sort of conjectural hypothetical message testing that has been used in polls and focus groups in getting to something that actually gets at the behavior of changing their
11:28am
mind. >> host: interesting. i mean to do this sort of work it does beg the question, how expensive is this? how much money do these campaigns have to raise to really do this as a level that you are talking about? ass it's expensive, but i think campaigns have invest intelligently and it will tell you that it should pay for itself. so one of the benefits of sort of microtargeting is unhealthy in the early process of sorting through the electorate. and so if you could cut 10 or 20% out of your persuasion universe because you are smart about who you should be talking to, plus 10 or 20% less of them male budget, phone budget, if it helps you by your tvs smarter, then you can either save money or redirecting resources. so there are campaigns that sort of my microtargeting adds a something that add-on to their
11:29am
male budget and in that case it's an expense. and it may help them with votes or not. campaigns sort of build certain types of analytics and targeting into the earliest strategic projections should become smarter about how they spend all the other money because they are not spending persuasion messages to people who already vote for them, or people will never vote for them. they are sending the right messages to people so it's not that will send every woman a message about abortion. we will send 18% people that we model actually movable on abortion messages and then figure out what the 82% rest of them get talk to about. so i think that, you know, that it's expensive to analyze and it certainly expensive as the obama campaign has done to build an in house shop with dozens of
11:30am
people, many of whom are working private sector jobs that paid good money before. but i think the campaign manager will tell you that it will allow them to be smart and get smarter, the more did he get in the more refined your models and predictions are. and ended and it pays for itself. >> host: i think we're going to take a break now and we will be right back. ..
11:31am
which for the last half-century has been tv ads and radio ads. those are very crudely and certainly via a market at time of day or certain channels, but you can't do a lot worse artist best the city than that. the other side of the committee keeshan is voter contact. talking to individual voters are you nobody went to joke about specific messages or get out the vote reminders to them. here, you can put together a campaign called the university are targets for something and because people are registered at a specific address and you have a number, you know who you're talking to them who you are not of the household level. some of the big challenges of any campaign going back in 200 years in the united case is
11:32am
figuring out, okay, tax people individuals. you are talking to? what are we talking to them about. we used to have your information on each voter. their name, age, address, gender and in some places the race. and then you had some information that she could sort of extrapolate based on where they live, which is what the senses told you about their tracker area in which he knew about their precinct, which is too new to legislation states that had it, but not everybody does exactly come assert the threat the 20 century we have a lot of people who are voting a certain elections and vice versa. the precincts, hundreds of people often gave some political or in haitian and so campaigns,
11:33am
absurdity computing power increased for the middle of the 20th century, it's sort of the way of dividing the elect are, this basic triage every campaign wants to do. you know, who are my people go turnout and support ramco from your ready because i'm not going to this dollar talking to them. with the people double support ipod because i don't exactly get me there. who's in the middle? once i got was the middle, search and figure out what i about your so precincts offered a very basic way of dividing a state or congressional district. or if you're a democrat, you would say, people and strong democratic precinct, i just want to turn them out because of his subject% democratic precinct committee but if i turn about people there, i'll turn out people who vote against me, but the margin is large enough that an aggregate will turn a as for myself. i'll ignore the republican
11:34am
precincts. and in the precincts of the middle, that are 40%, 60% have tried to chew that you individualize each voter to know who they're going to vote for and what issues they care about. and that is what people set up phone banks. setting up phone boxes managers can visit process in the 50s and 60s because interstate: is expensive, so you had to call the area and set up a phone. but as soon as you did, someone would start calling voters up of the voter rolls and say, who do you plan to vote for? one of things about the united states is people are used to telling strangers who they are going to vote for. so then you come up with a list of how individuals plan to vote and often a little bit of information about the issues they care about. in the same indoors. this is what the core of kansas is trained to identify voters. and then they start sorting
11:35am
people out here to do i have on my site already? who do i need to keep targeting with mail or persuasion phone calls? and so, this is the way campaigns have always worked spirit was changed in the starting and end of the last century coerce her to getting all the other data about people goes well beyond what the senses about their area of what their voter registration told you about themselves. it wasn't traditionally political information, but a lot was by commercial marketers are especially the credit rating agencies to which really had ventures in accumulating as much possible information about demographics, consumer habit, your lifestyle. and what happened starting in the late 1990s with people in politics discovered they could use what they know about you politically and what they know about your precinct with all these other sources of
11:36am
information and use statistical modeling techniques to basically run algorithms that could predict how you would answer the question of who are you voting for her. so instead of in a given state, having to set up phone banks or send out canvassers to knock in a million doors of the people in the precincts that are in the middle, you could use these modeling techniques to predict what each of those people would say. and then you do the same thing, which is the people in the middle persuadable, you still send them mail, but you know what an individual level you would write of people based on geography or just based on voter registration. and so, this technique really took hold 2002, 2004. the bush campaign was the first major camp into service successfully design and implement a system with this type of micro-targeting form of
11:37am
the campaign did. but it does this use tons of available individual level information, automate the process of predicting how someone will vote so you are no longer doing triage by precinct, but triage by individual and now i can take two people in neighboring apartments in without knocking on either of their doors, to predict who is likely to be with me, is likely because my opponent and who is undecided and come up with a separate treatment for each. >> the need for canvassers in traditional door-to-door and outside, has subsided? do we not think that it's much in much in these campaigns? >> i think campaigns like the obama campaign is smart about. you know, they do send people to do ids and those ids feed into their statistical models. in 2008, the obama campaign was
11:38am
basically modeling the electorate in swing states every week, which was to say that they would put out tens of thousands of payphone calls and have all the stuff for volunteers to talk to voters. that data and commercial vendors were feeding into these algorithms and every week they came up with the new projection of how each individual would vote, whether they were so, what issues they cared about. so that allows you to sort of guided canvassers to talk to people about particular issues. and so, there's less of a need to do all of your densification work individually. it helps to have volunteered to the feeding and that the more data points you have, the better predictions will be. one of the shifts will be using their canvassers far more for persuasion and not just for ids. historically campaigns usually thought of it as a use for
11:39am
volunteers to do the work of asking people who they're going to vote for and that lesser put together a mail list or if volunteers not than a thousand dollars and from 100 people turned abortion, august is names to the guy who does my mail pieces and how make abortion brochures and put them in their mailboxes. now, you can do a lot of sort of the predict dean, which people are most likely to be responsive to your message on abortion. and then maybe use your volunteers come especially like the obama campaign, you have a circuit of people willing to help and maybe you send them out to knock on doors or call a hundred people or have a real conversation with them. one of the things that has come out of this body of experimental work is that meaningful volunteer, person-to-person communication means something. it has an effect. it has an effect that paid call centers don't have. you know, it has, when callers
11:40am
have what they call chatty calls, which means they don't just go by script verbatim, but are able to set it improvise a little bit and sound more natural, people on the other hand clearly perceive that because it makes them are likely to turn up to the after they've had that conversation. to one of the themes running through this whole book is that this is all, you know, high-tech, modern style that is made possible by this massive collection of data taking place comment by advanced modeling techniques, computing power, all of the things 20 years ago in the form they do now. but there's been a kind of rude awakening of the human interpersonal dynamic in politics and i think we've seen a number of campaigns commenting history of the dean campaign and four, bush in 04, campaign to
11:41am
obama in at a weight that had a real volunteer culture. and i don't think those things are in conflict. and so, i think that campaigns are recognizing that because they're targeting is better than that because these person-to-person human interactions to make a difference that they can use -- that they can direct the time of their volunteers more effectively. so campaigns that are intelligent or integrating the two and a way that makes sense to do so supporting and not in conflict with one another. so i think that is what you'll see on sidewalks and i'll be up on soaping up is the campaign using advanced modeling techniques, lots of database and quite handsomely for. but it's on the service of figuring out who you get to volunteers to talk to and when and about what appeared that is one of the things that when i
11:42am
look back at the decade now. i've written about in the book, that i think is precisely not -- this is sort of a big brother moment in politics campaigns. they know more about you, no what moves you. in many ways we see sort of the opposite, which can then are also seeing the value and cannot measure the value of human interactions. smart campaigns want to figure out how to allocate those as intelligently as they allocate tv buys. >> personal touchstone matters. >> i think he used to be there was some mythology that a personal touch is better, the campaigns are really bad to send to stress their volunteers to talk to people. and so i think there was a certain type -- campaigns ever centralized to the obama campaign is essentially says any of them. since we are now able to measure the value of certain types of volunteer interactions of
11:43am
people, campaigns are going to concede a little bit of control and do not assume the message is always fast process and care to television ads or mailed. it is designed in a war room at headquarters, but that there is a benefit to having people talking to people, especially if you've done the difficult invisible work of telling which people to talk to and when and about what. so i look at the obama campaign in a way because they were quite good at molding the batch bran, right brain approach to modern campaigns and their ill to spend lots of money and advanced data and also lots of volunteers and field staff. i think working through that is one of the things in this era is most important to figuring out how campaigns can balance their need and desire for control, control their message from
11:44am
essentialist tactics and strategies, but also maximize the energy and talent and value of their their supporters doing names on their own to other voters. >> and you just prepare us? and implementers swing states for the election is decided, i would imagine all the stuff you talk about is focused like a laser in some of those states. could you just remind us but the state solder and a statue that most of these techniques are targeted? >> yeah, we're more or less than a dozen states now, week to week. you know, one thing important to rely is that not all swing states are equal or alike. we tend to see polling that's come to states where both candidates are between 46% and 49% in campaigns around the air
11:45am
and saying that their swing state, they are all sort of the same category. and campaigns are deciding where they compete and how they compete a stunned locals. this is the central strategic document of the campaign. when i read about the two 2008 campaign, david class would call this his bible, this binder is targeting us put together that has the vocals and for each state they would have to those they needed to land they left a bit of a buffer from a certain guarantee a margin of victory. but the question of how you get to that number in a given state is going to be different. and in some states, it was all mobilization. you know, in some states,
11:46am
pennsylvania has a good example of that. if democrats mobilize votes that they have -- supporters they have in the state, there's pat to victory to mobilization. states relied on persuasion. ohio is a pretty good example of that. and then you have states that relied on registration. so there's three numbers for each state. registration targets. in north carolina, the campaign if they had to register voters and strategically it may look as hourly as though i'm pretty sure they think differently about where they put their tv ads were the candidate goes.
11:47am
you know, i wrote in 2008 on the obama campaign and there was coming in now, one little example is john legend get a small concert and the town -- adam at the time is, but the way of saying the town is not a large city. he was in columbus and cleveland. i can't remember at the moment. the reason he went there was that they had seen registration numbers were lagging in this particular area and to reach their legislation goals on which they disaggregated down to this particular piece of turf, that they had him do a concert oriented towards registration, right man at the city hall or wherever people could go on register. it wasn't that they sent john legend and to turn people out.
11:48am
that is happening in broadway's everywhere. so there will be told states they talk about at the end of october, but each campaign will be competing differently based on their vocals, which are coming out of the marker targeting positions. they are basically taken every name of any individual person they think is very supporting and considering most people turn out targets and picking every name of people that they think of as persuadable and that is informing the tv buys, where the candidate goes come at the, everything. >> if you were to say, with everything you know, what do you think that key state -- is there one key state that it comes down to? are one or two you're particularly interested in for this cycle that intrigues you? >> i mean, virginia and north carolina are intriguing to me
11:49am
just because they are new states to presidential politics. and so, i think it's interesting to see how the campaigns that don't necessarily have a sort of reflect this attack takes work in eating and then adjust for the parties are not equally established and are used to run in presidential campaigns. the one thing generally, after spending a year or so talking to decide seaplane campaigns in some of the people who wrote the most granular familiarity with not just data and some of the statistical bottles about politics, but also the real empirical research of cause and effect is that i put us in the state and people who are confident about this they know or their predictions because some of the smartest people i talk about are the ones who are the most willing to say about very big names.
11:50am
i don't know that. i don't think it's knowable. there's a lot in campaign -- there's a lot of there's a lot of it in this took the relentless 1015 years for a vc when politics have learned about what works and what doesn't. but it is also meant those people far more aware of the limitations of knowledge. and those limitations are significant. there is a massive amount about what happens in that election that is unknown and at the moment or less unknowable. anybody that speaks of too much confidence about how things are or how they will be probably sent trusted. >> we've heard a lot about mail another tactics, but tv advertising. i mean, what does that whole world look like as we see more of the internet and tv coming together and be one of the same? can you paint a picture?
11:51am
>> guest: gas, two separate things are happening. the basic question of knowing who they are showing arrived to when you buy an ad on the broadcast or cable. we are getting better at that. one thing that's interesting, i write about rick perry's 2006 gubernatorial reelection and it was the first campaign to do large-scale randomized six. one of the experiments was a randomly assigned as tv had for three weeks across states a media market. one of the professors, or these four academics were called to eggheads by perry's advisor, dave carney. one of them is professor wright university of maryland, relies for talk about media market, but do we really know -- media market shoes were a signal for a tv in the. anybody driven at night and try to listen to a ballgame, or if you can hear some foreign city that you are not in, the station
11:52am
realizes that almost certainly the balance of the media market or not the neat circle around the city that were used to dealing with. so he ended up partnering with the graduate student, and the naval academy, they started came again, which is an engineering technique to measure where the signal goes. and it turns out the edges of the media market look absolutely nothing like what we generally think of as media markets. so much of our conversation is how many points did you put into which media market? well, turns out the media market doesn't look like people were inside or outside are your thoughts were on the other side of the line. it's not consistent. the other thing is points on campaigns realize they are really crude measurement. 100 points is simply the average number of times a given person
11:53am
cheesier i'd. people have john -- there's a few times where nielsen is basically allowed political characters to match individuals. their individual viewing habits to a voter file. an model individual things. the terms i guess it can average. the average distributes and a crazy courtesy think when you think about your friend's tv viewing habits, yes. so the average may be that each person sees at five times can a handful of see it 20 times. a lot of people see it once. very few people actually see it five times. and this has been the way the tv ads have been bought for a generation now. on one end, campaigns are a lot smarter at understanding when you buy a point of tv who is actually being exposed. they're able to make much
11:54am
smarter, less crude decisions about where the broadcast ble followers go because of that. the other thing is fewer people are watching tv via television. the natural substitute that is web advertising because you can still run video. video is obviously an amazingly valuable media that can't be replaced by mail or by phone. i mean, voters like to see people want to lead them to do a lot of things with motion picture. and obviously you can run video ads online. the big challenge, and we see it dances every month now is knowing as much about who the lower online as we do about who they are offline. in still, you know, politics takes place offline. people vote at their home address. and republicans vote in
11:55am
republican primaries away novus republican at each individual house. the challenge is making sure you know who those republicans are in which district they lived when they are surfing the web or watching videos on hulu. so the challenge now and campaigns are doing much better about.com is bringing some level of geographical political targeting to web advertising, whether it's video or simple banner ads. you know, intel this year, the communication is still primarily used as a fundraising out let. most of the reason was that you can measure rather affect the playbook works in terms of fundraising online. you run a web ad and you can get a very good measure of fear response rate and figure out who is donating, who's not, what amounts get what they respond
11:56am
to. we are now starting to get an update to use it as a persuasion and geo tv to all. that will allow campaigns to have a better idea of who they should talk to online, who they should talk to unconventional tv. >> i know we have a couple minutes. i want to ask tumor questions. privacy, should the voter be concerned campaigns are getting more and more data and using this data? >> what they've always had kind of these casually asserted less of an of privacy in politics than we have in other areas. and so, it's worth remembering bobby for the internet came on, were always pretty condition, even if we found it frustrating to get the call to dinner from someone who you knew your name, party registration maybe they tell you one of my colleagues called you last week into two about abortion. i'm calling to remind you that
11:57am
she talked to my colleague about abortion and can rescind a ban to your house next tuesday morning and drive you to the polls. those are invited to came during dinner, but no one found it no one found it that invasive. so fundamentally the campaign do with a lot more data in new method isn't a lot we are not. your record as a voter is public and whether or not you're registered, campaigns collect that information. one thing ever written about this year is at least on the internet, partially because campaigns are really afraid of the backlash of getting lumped in with more pressing things to consumer marketers are doing, that they are bending over backwards to respect certain types of privacy restrictions on cookies online that they are not required to end his eon the legal standard. so i think that on the flipside of the campaigns know more about
11:58am
you, but in many ways, p. m. keynes usage they saw to to buy the things they want to talk to you about. campaigns have no interest in doing anything other than catering to your individual needs or interests and that is that they're doing. >> we have a minute or two before we close. but what you want to say to people listening, learning about this. we are weeks away from this election. but should they luck out for bikes what advice can you get to the voter in terms of what they should eat paying attention to, particularly peeved on the swing states. >> guest: one of the things that comes out of congress on this is that things that you don't get necessarily as being the high side can paint or the places were the biggest innovations in the sophisticated research and canvassing scripts. you now, the kid knocks on your door, asks you a few questions.
11:59am
amazingly, that is the place where i must greatest energy and insights from behavioral psychology has been brought to bear. so i would listen really closely to what people ask you on the phone, what they ask you at the door, why your mail looks the way it does because my guess is that a lot of the misinformed by an amount of sophistication about the human mind that would shock you. >> host: well, thanks so much. "the victory lab: the secret science of winning campaigns," this is just the time i him is so provocative and seems like it when the groundwork for things to come. so i just want to say thank you, sasha, for doing this. are you spending a lot of time in the road looking at mark and dave in this condensed. time now? >> guest: i've been campaigning for slate on this,